Despite newer treatment modalities, overall mortality from AML still exceeds 50 percent. Differentiation agents have shown promise but due to lack of efficacy and toxicity we are still looking for additional or better agents. In the initial studies, G-CSF could override the differentiation block in leukemic cell lines and in some fresh AML cells. However, most responsive AML cells proliferated without differentiation upon exposure to G-CSF. Based on our recent studies we postulate that in AML cells, Stat3alpha activation may interfere with the G-CSF-driven differentiation signal. In this proposal, we plan to determine 1) if Stat3 is required for G-CSF-driven myeloid differentiation, and 2) the effect of altered relative levels of Stat3 isoforms Stat3alpha and Stat3beta on G-CSF-driven myeloid differentiation. By meals of antisense inhibition of Stat3 and use of a dominant negative of Stat3, we shall confirm the role of Stat3 in myelopoiesis. Then by overexpressing Stat3alpha, we shall examine the role of relative Stat3 isoforms level in G-CSF-driven myelopoiesis. The long-term goal of this project is to apply information gained from these studies to the design of new differentiation-inducing agents for the treatment of AML. Such therapies might target Stat3alpha at the level of its activation or expression in AML cells.